CN1922463A

CN1922463A - Method and apparatus for measuring film thickness by means of coupled eddy sensors

Info

Publication number: CN1922463A
Application number: CNA2004800421043A
Authority: CN
Inventors: Y·戈特基斯; R·基斯特勒; A·欧萨斯; C·弗罗因德
Original assignee: Lam Research Corp
Current assignee: Applied Materials Inc
Priority date: 2003-12-30
Filing date: 2004-12-06
Publication date: 2007-02-28
Anticipated expiration: 2024-12-06
Also published as: JP2007517218A; CN101524829B; CN100507437C; KR20060116847A; US7205166B2; EP1706704A1; JP4590416B2; US20040206455A1; CN101524829A; WO2005066579A1

Abstract

A method for minimizing measuring spot size and noise during film thickness measurement is provided. The method initiates with locating a first eddy current sensor directed toward a first surface associated with a conductive film. The method includes locating. a second eddy current sensor directed toward a second surface associated with the conductive film. The first and second eddy current sensors may share a common axis or be offset from each other. The method further includes alternating power supplied to the first eddy current sensor and the second eddy current sensor. In one aspect of the invention, a delay time is incorporated between switching power between the first eddy current sensor and the second eddy current sensor. The method also includes calculating the film thickness measurement based on a combination of signals from the first eddy current sensor and the second eddy current sensor. An apparatus and a system are also provided.

Description

Method and apparatus by coupling eddy current sensors MEASUREMENTS OF THIN thickness

Background of invention

Relate generally to semiconductor of the present invention is made, and more particularly, relates to the on-line metering that is used for the wafer processing procedure process control and learns.

A plurality of steps are arranged, wherein the formation and the removal of the various layers of base substrate experience in semiconductor fabrication.Little characteristic dimension, strict surface plane requirement combine with constantly seeking productivity gain, make people extremely wish when reaching correct thickness, that is, when reaching the terminal point of processing step, stop described process.

Eddy current sensor is used for the measurement of displacement, the degree of approach and film thickness.Described sensor relies on the induction by current that is caused by the 7 fluctuation electromagnetic fields of pressing close to the detecting coil of just measured object on sample.The fluctuation electromagnetic field is owing to exchange current is set up by described coil.The fluctuation electromagnetic field eddy current of inducting, eddy current is upset main field, consequently changes the inductance of described coil.

Fig. 1 is the simplification schematic diagram of eddy current sensor principle of work.Exchange current flows through the coil 100 of conductive body 102.The electromagnetic field of the described coil eddy current 104 of in conductive body 102, inducting.The amplitude of eddy current and phase place influence the load on the described coil again.Thereby, the influence of the eddy current that near the conductive body the impedance of described coil just is subjected to being arranged in is inducted.Measure this influence, so that detect the degree of approach of conductive body 102 and the thickness of described object.Distance 106 influences the effect of 104 pairs of coils 100 of eddy current, therefore, if object 1002 moves, then also will change from the signal of monitoring eddy current to the sensor of coil influences.

Attempt to use eddy current sensor to come the effect of thickness of MEASUREMENTS OF THIN limited.Because from the signal of eddy current sensor to the thickness of film and substrate to sensor apart from both's sensitivity, must find the solution two unknown numbers.Fig. 2 is the schematic diagram with chip carrier of eddy current sensor, and eddy current sensor is used for measuring the thickness of wafer in chemical mechanical planarization process (CMP) process.Chip carrier 108 comprises eddy current sensor 110.In the CMP operating process, the wafer of being supported by the carrier thin film 112 of carrier 108 114 is crushed on the liner 116, so that with the surface planarization of wafer.Liner 116 is supported by stainless steel base plate 118.

A shortcoming of Fig. 2 configuration may change+/-3 mils from the changeability of carrier thin film.Thereby carrier thin film causes sizable changeability of distance between wafer and the sensor.In addition, the different downward pressure that is applied on the carrier thin film causes further variation along with the compression of carrier thin film.Therefore, to influence distance and the latter influences all these variablees of sensor thickness measured value and demarcates, become extremely difficult.Another shortcoming of this configuration is to cause that by the existence with another conductive material that is just separating at measured conductive material this generally is called the 3rd object effect.If the thickness of conductive layer is less than so-called skin depth, then the electromagnetic field from coil will can not absorbed fully, and will be partly through the stainless steel base plate 118 of Fig. 2 liner 116.This additional eddy current of will in stainless steel band, inducting, thus influence is from the resultant signal of eddy current sensor.In addition, should be pointed out that liner along with the wearing and tearing of time or corroded the variation that causes distance between stainless steel base plate and the eddy current sensor, this influence is to the contribution of eddy current sensor resultant signal.Thereby, when handling wafer continuously, must consider wear factor.Thereby owing to be introduced into the variable cause of thickness measure, the quantity of error is high as can't to accept and unpredictable.

In view of the premises, be necessary to eliminate or compensate in working order intrinsic changeability down, so that can determine terminal point accurately, reach required thickness more accurately.

Brief summary of the invention

In the broadest sense, the present invention is by determining under desirable state that is at the thickness of idle state lower wafer, and provides described thickness so that can consider or offset the changeability that is caused by the unknown number of introducing in the process operation process, satisfies these demands.

According to one embodiment of the invention, provide a kind of being used for that the noise of measured film thickness process is reduced to minimum method.Described method is from being arranged to first eddy current sensor towards the first surface related with conductive film.Described method comprises is arranged to be positioned at an opposite side of conductive film to second eddy current sensor and towards the second surface related with conductive film.First and second eddy current sensors can sharing of common axis or depart from each other.Described method also comprises alternately to first eddy current sensor and the power supply of second eddy current sensor, makes win eddy current sensor and second eddy current sensor once have only one to be powered.In one aspect of the invention, between the operation of Switching power between first eddy current sensor and second eddy current sensor, comprise time delay.Described method also comprises the measured value of basis from the combination calculation film thickness of the signal of first eddy current sensor and second eddy current sensor.

In another embodiment, be provided for shining upon the sensor array of wafer thickness.Sensor array comprises a plurality of top sensor and in a plurality of bottom sensor of the opposite one side of top sensor.In in a plurality of bottom sensor each and a plurality of top sensor corresponding one coaxial.Also a plurality of bottom sensor be configured to when in a plurality of top sensor corresponding one inoperative when working.Comprise the power supply of not only keeping in touch but also keeping in touch with a plurality of bottom sensor with a plurality of top sensor.Also comprise and being configured to from described power supply alternately to the controller of a plurality of bottom sensor and a plurality of top sensor power supply.

According to another embodiment of the present invention, provide a kind of system that is used to handle wafer.Described system comprises chemical-mechanical planarization (CMP) instrument.The CMP instrument comprises the chip carrier that is limited in the shell.Chip carrier has the bottom surface, and the bottom surface has definition window thereon.Carrier thin film is attached to the bottom surface of chip carrier.Carrier thin film is configured to support wafer in the CMP operating process.Sensor is embedded in the wafer.Sensor is arranged on the end face of described window.The described sensor configuration eddy current that becomes to induct in wafer is so that determine the degree of approach and the thickness of wafer.Described sensor array is arranged on beyond described CMP instrument.Sensor array and the sensor communication that embeds chip carrier.Sensor array comprises first sensor and corresponding second sensor.First sensor and corresponding second sensor configuration become in action state and inoperative state between alternately.First sensor also is configured to be in the state that works when second sensor is in inoperative state.Sensor array is configured to detect and first sensor and corresponding second sensor wafer thickness signal to the range-independence of wafer.

Understand, as state that above general description and following detailed description are exemplary and only are exemplary rather than for restriction of the present invention.

The cutline of accompanying drawing

Be included in this instructions and constitute its a part of accompanying drawing graphic extension one exemplary embodiment of the present invention and be used for explaining principle of the present invention with described description.

Fig. 1 is the simplification schematic diagram of the principle of work of eddy current sensor.

Fig. 2 is the synoptic diagram of chip carrier, and chip carrier has eddy current sensor, is used for measuring in chemical mechanical planarization process (CMP) process the thickness of wafer.

Fig. 3 is the simplification schematic diagram of coupling sensor, is used for the thickness of the wafer that enters according to one embodiment of the present of invention measurement.

Fig. 4 is the curve map as the signal of the coupling eddy current sensors that disposes in Fig. 3 according to one embodiment of the invention.

Fig. 5 is the simplification schematic diagram according to alternative configuration of the coupling sensor of the thickness that is used to measure the wafer that enters or film of one embodiment of the invention.

Fig. 6 A is the simplification schematic diagram according to another kind of alternative configuration of the coupling sensor of the thickness that is used to measure the wafer that enters of one embodiment of the invention.

Fig. 6 B is a curve map, graphic extension according to one embodiment of the present of invention when utilize described curve map one side configuration be used to detect the eddy current sensor array of film thickness the time, the stability of average signal.

Fig. 7 A is the simplification schematic diagram that enters thickness transducer according to the CMP process that the is coupled to downstream of one embodiment of the invention.

Fig. 7 B is the simplification schematic diagram that is coupled to the alternative embodiment that enters thickness transducer of downstream CMP process thickness transducer.

Fig. 8 A and 8B are curve maps, graphic extension according to one embodiment of the invention from the signal of the eddy current sensor that is used for film thickness with from the correlation between signals of measuring resistance rate measured film thickness device.

Fig. 9 is a curve map, and graphic extension is according to the calibration curve of eddy current sensor that is used to measure the copper film thickness of one embodiment of the invention.

Figure 10 is a curve map, and graphic extension is according to the output signal of two eddy current sensors in the CMP operating process of copper film on substrate of one embodiment of the invention.

Figure 11 A is the curve map according to infrared ray (IR) sensor signal of the temperature time to time change of the measurement sand belt of one embodiment of the invention.

Figure 11 B is the curve map of the infrared signal that changes along with the time of the chip temperature according to one embodiment of the invention.

Figure 12 is the synoptic diagram of 30 seconds time series T1-T9 of CMP process, and graphic extension is according to the removal by the copper film of the vortex sensor measuring in the chip carrier of one embodiment of the invention.

Figure 13 simplifies schematic diagram, graphic extension according to one embodiment of the invention be applied to or skew or coaxial two sensors on the dutycycle of handover scheme.

Figure 14 A and 14B are exemplary graph, and graphic extension is poor according to the non-switching power supply plan (powerizing scheme) and the noise between the switching power supply plan of one embodiment of the invention.

Figure 15 A and 15B are exemplary graph, are illustrated in the noise that suffers in the non-switching power supply plan.

Figure 16 A and 16B are exemplary graph, and expression is similar to the reading of Figure 15 A and 15B when the switching power supply plan is applied to sensor according to one embodiment of the invention.

Figure 17 simplifies schematic diagram, is illustrated in according to the edge exclusion in the arranged coaxial of one embodiment of the invention and improves.

Figure 18 is a process flow diagram, and graphic extension is according to the method operation that is used for the noise of measured film thickness process is reduced to minimum of one embodiment of the invention.

DETAILED DESCRIPTION OF THE PREFERRED

Now describe several one exemplary embodiment of the present invention with reference to the accompanying drawings in detail.Fig. 1 and 2 discussed in " background of invention " part in the above.

Eddy current sensor (ECS) is used to measure the thickness of metal film of mobile wafer.Determine that ECS can provide enough response fast for the wafer that moves with typical loader device people speed.Therefore, might " in operational process " finish thickness measurements, and don't influence processing output.In addition, the motion of wafer can be utilized to a limited number of sensor generation thickness distribution curve from group configuration.For example, wafer aligner provides the motion of sense of rotation and linear radial direction.Therefore, the sensor group can catch the thickness distribution curve of wafer when wafer experiences public robotization processing of wafers scheme.In one embodiment, can produce the thickness distribution curve of each wafer, make the method method of downstream process to be optimized at thickness distribution curve.

Each accompanying drawing that starts below provides the sensor that enters or sensor group's configuration, and it is determined under the state of non-processing, that is the thickness of noiseless state lower wafer.Then, can store this thickness and/or with the downstream manufacture process of its notifier processes metallic film.In other words, the thickness that enters can be offered the similar measurer for thickness of downstream manufacture process.Then, can demarcate with the described thickness that enters, so that make the unknown number or the variable that cause by processing conditions become incoherent with the sensor that the downstream manufacture process interrelates.Should be pointed out that the sensor group is meant the combination of two or 3 sensors, be used to define new metrology characteristic.In one embodiment, sensor is combined among the group so that eliminate the susceptibility of adjusting the distance.The sampled result of the data of utilizing described sensor configuration collection here also is provided.The array of sensor used herein comprises two or more sensors group of the bigger area that is combined into covering (that is mapping) wafer.In addition, 3 sensors that are positioned at wafer the same side can be thought an array.

When the sensor of trooping is combined with the station that pre-exists, just can be the process operation mapping in downstream and the thickness of memory chip.In addition, when from process chamber taking-up wafer, can also scan handled wafer thickness, so that the feedback about the result of process operation is provided.Correspondingly, can adjust method of operating according to described feedback.Certainly, also can provide the aftertreatment result for further process operation.

Fig. 3 is the simplification schematic diagram according to the coupling sensor of the thickness that is used to measure the wafer that enters of one embodiment of the invention.Top sensor 130 and bottom sensor 132 are configured to provide the signal of the thickness of representing wafer 138.In one embodiment,

sensor

130 and 132 is eddy current sensors.Wafer 138 comprises substrate 142 and metal level 140.The axis 134 of top sensor departs from the axis 136 of bottom sensor 132.Those skilled in the art will recognize that, by skew top sensor 130 and bottom sensor 132 (wherein two sensors all are eddy current sensors), just can be not interfering with each other by

sensor

130 and 132 electromagnetic fields that produce, that is can not suppress each other.Frequency is identical and the place aimed at axis 136 of axis 134 therein, may occur the inhibition to these signals in some cases, still, will illustrate as following, can adjust, so that avoid inhibition to these signals.Controller 144 is communicated by letter with 134 with sensor 132.In one embodiment, controller 144 is configured to from the thickness of sensor 132 and 134 received signals and definite wafer 138.Should be pointed out that controller 144 asks the mean value from the signal of

sensor

130 and 132, so that obtain representing the signal of wafer thickness.In addition, by top sensor 130 and bottom sensor 132 are set, the variation of the distance 148 between top sensor and wafer 138 end faces between the bottom surface of the variation of distance 146 or bottom sensor and wafer 138 has been cancelled.In other words, when two sensors were all static, the variation by average these readings have been offset any one distance made that present described signal is the function of thickness rather than the thickness and the degree of approach.In another embodiment, controller 144 such as the CMP instrument, makes that downstream process can be according to the setting of the wafer thickness optimizing process that enters, such as downforce that is applied and tape speed to the instrument in the thickness notice downstream of calculating.

Fig. 4 is the curve map from the signal of the eddy current sensor that disposes in Fig. 3 of picture according to one embodiment of the invention.Output (volt) and time relation that curve Figure 150 draws eddy current sensor.Curve 152 representative is from the sensor that is positioned at below the wafer, such as the signal of the bottom sensor 132 of Fig. 3.Curve 154 representative is from the sensor that is positioned at above the wafer, such as the signal of the top sensor 130 of Fig. 3.Bold curve 156 is represented the mean value of curve 152 and 154.For those skilled in the art, curve 156 obviously provides a kind of signal of substantial constant.In other words, from the signal of top sensor with from the signal of bottom sensor, just got rid of the dependence that the signal of sensor is adjusted the distance by combination.For example, the wafer 138 along with Fig. 3 shifts near top sensor 130, described signal increase.The zone 158 illustrations the motion of wafer near top sensor.Thereby, correspondingly increase from the signal intensity of top sensor, as curve 152 graphic extensions in the zone 158.Simultaneously, wafer moves apart bottom sensor.Therefore, reduce, be similar to increase, as curve 154 graphic extensions in the zone 158 from the signal intensity of top sensor from the signal intensity of bottom sensor.Thereby it is constant that the mean value of top and bottom signal intensity keeps.Because signal intensity and sensor to the distance of object is a linear dependence, so object is to the caused change in signal strength of the motion of static first sensor, the inverse variation that is moved apart the caused intensity of motion of the second static sensor by object is offseted.This has eliminated the susceptibility that signal intensity is adjusted the distance again.

Thereby, by trooping sensor or have the sensor group (such as the sensor in the chip carrier that embeds the CMP instrument) of the sensor that works down in working order, can demarcate the sensor that works down in working order, so that provide more accurately about removing the information of speed and course end.In other words, provide the accurate measured value of the film thickness that enters or the thickness of wafer, make the error that under processing conditions, causes to compensate by caliberating device.In addition, described sensor group can be as continuous operational process control, wherein wafer has thickness distribution curve (by the thickness distribution curve of first sensor or sensor group mapping) and described distribution curve and is downloaded controller to process tool, so that customize described process, such as the CMP process, so that remove the correct number of film thickness.

Fig. 5 is the simplification schematic diagram according to alternative configuration of the coupling sensor of the thickness that is used to measure the wafer that enters or film of one embodiment of the invention.Here top sensor 130 and bottom sensor 132 are shared same Z-axis 160.In order to eliminate any interference or the inhibition of signal between top and the bottom sensor, can apply different frequencies in the sensor separately each.In addition, can apply phase shift, make two sensor phase differences.In other words, a sensor turns over turnback to signal waveform, so that eliminate the inhibition to signal.As described above, because got rid of distance as variable, so signal intensity is the function of thickness.Be expressed as mathematical equation: S=k (THK), wherein S is a signal intensity, and k is a sensitivity coefficient, and THK is a thickness.In one embodiment, be known place in signal intensity and sensitivity coefficient for aforesaid equation, can determine described thickness by calibration curve.Can offer the downstream process instrument of handling metallic film in the semiconductor fabrication to this thickness of determining,, be discussed as reference Fig. 8 and 9 such as the CMP instrument.In addition, can use the switching power supply plan that discusses in more detail with reference to Figure 13-16B.Therefore, can be two

sensors

130 and 132 power supplies with single power supply 145.Certainly, can be used to switch power supply plan to single power supply, wherein be offset these sensors as shown in Figure 3.

Fig. 6 A is the simplification schematic diagram according to another alternative configuration of the coupling sensor of the thickness that is used to measure the wafer that enters of one embodiment of the invention.In described embodiment, the sensor flock mating is set to determines the thickness of wafer 138 along the axis 162 of top sensor 130.

Location bottom sensor

132a and 132b like this makes that from each sensor to axis 162 distance equates.Thereby, by asking mean value, just determine described signal from the signal of

sensor

132a and 132b, thereby, determine thickness along axis 162.Here, do not relate to the interference or the inhibition of signal between top and the bottom sensor, because

bottom sensor

132a and 132b depart from the axis of top sensor 130.The motion of zone 164,166,168,170 and 172 expressions wafer 138 when wafer passes the space that defines between top sensor 130 and bottom sensor 132a and the 132b.With reference to Fig. 6 B these regional meanings are described in more detail.

Fig. 6 B is a curve map, graphic extension according to one embodiment of the invention when utilizing the stability of eddy current sensor group time average signal that is used to detect film thickness be configured in a side of described curve map.In described configuration, the eddy current sensor group comprises top sensor 130 and two bottom sensor 132a and 132b.Wafer 138 is walked between top sensor 130 and bottom sensor 132a and 132b.The draw relation of ECS reading (volt) and time (millisecond) of curve Figure 171.The reading of curve 169 representative sensor 132a, and the reading of curve 173 representative sensor 132b.Curve 175 is represented the reading of top sensor 130.In one embodiment, ask mean value from the signal of sensor 132a and 132b.Ask described result and mean value then from the signal of sensor 130.Last average signal is by curve 177 representatives.Curve Figure 171 graphic extension wafer 138 passes the different position of sensor group time.For example, wafer 138 enters the sensor group and is represented by zone 164 on curve map.Should be pointed out that 11 volts the ECS reading representative starting point corresponding with 0 thickness because wafer here also measure less than.Along with Waffer edge passes the sensor group, described signal stabilization gets off.Then, wafer continues to move ahead along this central walking path, expresses as zone 166.As can be seen, average reading keeps relative stability with curve 177 representatives.

Then, move 0.020 inch on the wafer.Although become stronger from the signal intensity of top sensor 130 (that is, the difference between the reference signal at the signal of actual measurement and thickness 0 place), become from the signal intensity of

bottom sensor

132a and 132b a little less than.But, still constant relatively with the mean value that curve 177 is expressed.Then, wafer moves down 0.020 inch from described central path.Correspondingly, become from the signal intensity of top sensor 130 a little less than, and become stronger from the signal intensity of

bottom sensor

132a and 132b simultaneously.As mentioned above, the mean value of top sensor signal and bottom sensor signal is still relative constant.Thereby as mentioned above, the sensor group provides stable signal, and it is irrelevant that wherein wafer leaves the distance of sensor, because the average wafer of having offset of top and bottom signal moves the changeability that causes, perhaps even the signal change that causes of chip warpage.Then, wafer 138 shifts out the sensor group and describes as zone 172, and signal changes herein, because sensor is seen the edge of wafer when withdrawing from.Should be pointed out that Fig. 6 a-quadrant 164-172 representative as with reference to the described similar motor pattern of Fig. 6 B.The movement profiles of Fig. 6 A produces the average signal shown in Fig. 6 B of substantial constant.Those skilled in the art will recognize that, sensor can dispose with the following methods: bottom sensor and top sensor, bottom sensor and two bottom sensor or any other suitable configuration, offset the motion of wafer so that keep stable reading.

Fig. 7 A is the simplification schematic diagram that enters thickness transducer that is coupled to downstream CMP process thickness transducer according to one embodiment of the invention.Here, enter the thickness of the film 140 above the substrate 142 that thickness transducer 130a determines the thickness of wafer and/or wafer 138.The signal conveys of the described definite thickness of expression to controller 144.Controller 144 is again the sensor 130b of described signal conveys to the chip carrier 174 that is used to embed the CMP process.In one embodiment,

sensor

130a and 130b are eddy current sensors.In another embodiment,

sensor

130a and 130b are infrared ray sensors.Should be pointed out that by the thickness of the wafer 138 that enters is provided to sensor 130b, just can finish demarcation, so that eliminate susceptibility basically distance between sensor and the wafer.The changeability of distance can be caused or only caused by the intrinsic change of carrier film thickness by carrier thin film during the various duties 176 compression between sensor 130b and the wafer 138, it can up to+/-3 millimeters.In addition, the distance affects between polishing underlay 178 tops and the stainless steel base plate 180 is from the signal of sensor 130b.In addition, the signal of the thickness of the wafer 138 that enters of expression can be used for calibration sensor 130b, so that eliminate the polishing underlay franchise basically and liner corrodes caused changeability, it influences distance between polishing underlay 178 tops and the stainless steel base plate 180.

Fig. 7 B is the simplification schematic diagram of the alternative embodiment that enters thickness transducer that is coupled to the thickness transducer of downstream CMP process.In Fig. 7 B, communicate by letter with controller 144 with the sensor group that 132b constitutes by top sensor 130a and bottom sensor 132a.Here, suitable sensor group is set, such as the sensor group of Fig. 6 A and 6B, so that determine the thickness that enters of film 140 wafer 138 or described wafer.Those skilled in the art will recognize that, also is the suitable sensor group that can be used for determining wafer 138 thickness with reference to the sensor group of Fig. 3 and 5.In one embodiment, controller 144 is asked the mean value from the signal of

bottom sensor

132a and 132b, so as to determine about wafer 138 along the axis of top sensor 130a thickness signal by the thickness of wafer.Then, ask bottom sensor average signal and mean value, so that determine the thickness of wafer 138 or film 140 from the signal of top sensor 130a.Then this thickness is transported to and embeds sensor 130b.As top pointed with reference to Fig. 7 A, can demarcate automatically sensor 130b, wherein get rid of basically to the susceptibility of distance between these sensors and the wafer 138 with to the susceptibility of distance between the top of polishing underlay 178 and the stainless steel base plate 180.In other words, can demarcate automatically in real time, so that the caused ECS reading of variation of the sensor degree of approach that liner wearing and tearing or CMP carrier to other mechanical drift problem of the mechanical shift of plate of adjusting causes.

The sensor 130b of Fig. 7 B is arranged on above the pad 175.Pad 175 is aimed at the bottom surface of chip carrier 174.Pad 175 is made up of any suitable non-conducting material.In one embodiment, pad 175 is polymkeric substance.In another embodiment, the thickness of pad 175 is between about 1 millimeter (mm) and about 1.5 millimeters.Should be pointed out that pad 175 for sensor 130b provides window, be used to send and receive expression wafer or the thickness of wafer upper film and the signal of the degree of approach.

Although the embodiment graphic extension of Fig. 7 A and 7B is used for sensor or the sensor group of CMP before handling, also can be positioned at sensor or sensor group after CMP handles, be configured to improve the information that continuous operational process is controlled so that provide.Although enter thickness for specific method is downloaded to the process tool station so that compensating any film thickness that enters gets ready, the thickness after the CMP is for proofreading and correct any process variations that detects of determining in the thickness evenness measurement after CMP.In other words, provide CMP thickness evenness measurement result afterwards, so that the further fine tuning of sensor 130b is demarcated be provided with, to obtain terminal point accurately as feedback to sensor 130b.In one embodiment, controller 144 provides feedback from back CMP sensor group to sensor 130b.In another embodiment, eddy current sensor is the eddy current sensor that generally can obtain, such as the GP-A series analog displacement transducer that can obtain from SUNX Limited (company).In another embodiment, a plurality of sensors can be set in the chip carrier of Fig. 7 A and 7B.Described a plurality of sensor can be linked at together, so that detect the degree of approach and the thickness of metal film of wafer.For example, can be included in capacitive transducer in the chip carrier, so that the distance between definite wafer and the ECS sensor.When capacitive transducer is linked to the ECS sensor, can provide described distance to the ECS sensor.

Fig. 8 A and 8B are curve maps, and graphic extension is according to one embodiment of the invention, from the signal of the vortex current sensor of MEASUREMENTS OF THIN thickness with from the correlation between signals of measuring resistance rate measured film thickness device.Curve 190 representatives of Fig. 8 A are from typical electric resistance sensor method signal.There is the 3rd metal object in curve 192 representative, under the situation such as the stainless steel base plate of the polishing underlay of CMP instrument or band, chip carrier, air bearing platen etc., at the signal that leaves on the different point of center wafer from vortex current sensor.Curve 194a representative is from the signal of the eddy current sensor that does not have the 3rd metal object.Thereby eddy current sensor is closely related with signal from standard four point methods of resistance measurement.In addition, percentage deviation+/-5% in, as triangle 196 graphic extensions, wherein each triangle is corresponding to the difference between the corresponding each point on each curve of described each triangle top.

Graphic extension and vortex sensor measuring value like Fig. 8 category-B from the signal correction of typical resistances sensor, method.Here, curve 190b representative is from the signal of typical electric resistance sensor method, and 194b representative simultaneously is from the signal of the eddy current sensor that does not have the 3rd metal object.Should be pointed out that the wafer of measuring is different from the wafer of measuring in Fig. 8 B in Fig. 8 A.ECS signal 194b also is closely related with the RS75 signal.In other words, the percentage deviation between these signals generally+/-5% within, as triangle 196 graphic extensions.Those skilled in the art will recognize that, the point at these curve two ends is corresponding to the edge (that is edge of wafer) of film, thereby do not think that the point at these curve two ends is correlated with.

Fig. 9 is a curve map, and graphic extension is according to the calibration curve of eddy current sensor that is used to measure the copper film thickness of one embodiment of the invention.The film thickness of copper and the ECS voltage readings that is associated of described thickness when there is not the 3rd metal object in curve 198 representatives.Curve 200 is represented under the situation that has the 3rd metal object, the ECS voltage readings that is associated of the film thickness of copper and described thickness.Those skilled in the art will recognize that described calibration curve can be applied to top with reference to the described sensor of Fig. 3,5,6A, 6B, 7A and 7B.In addition, calibration curve can produce at any conductive membrane layer, describe copper here just for the ease of graphic extension, and and do not mean that restriction.

Figure 10 is a curve map, and graphic extension is according to the output signal of two eddy current sensors in the CMP operating process of copper film on substrate of one embodiment of the invention.The time dependent signal of forward position ECS signal of the wafer of curve 210 representative experience CMP operations.The wafer of curve 212 representative experience CMP operation back along time dependent ECS signal.Should be pointed out that the zone that is defined between curve 214 and 216 removed the surface characteristics of wafer (topography), and copper capping layer has been removed from wafer in the zone after the curve 216 simultaneously.Point 210-1 represents terminal point, that is, remove the copper capping layer (overburden) in described wafer forward position.Point 212-1 represents the terminal point on edge behind the wafer.Those skilled in the art will recognize that the information of gathering from the ECS that embeds chip carrier will produce continuous data, be used for determining the speed of removal.In addition, can observe the change of removing speed between forward position and the edge, back.Embed place in the chip carrier at sensor, sensor provides about just in the continuous real time data of the thickness of the thickness of measured wafer or the film on the wafer.In other words, do not exist snapshot of the every circulation primary sensor of sand belt or liner picked-up that the window of discrete measured value is provided.The embodiments described herein provides successive measurements and thickness monitor.

Figure 11 A is the curve map according to the time dependent infrared ray of temperature (IR) sensor signal of the measurement sand belt of one embodiment of the invention.Those skilled in the art will recognize that silicon substrate is transparent to infrared signal, and therefore, infrared signal just can detect the temperature at the film of the wafer of the surface planarization on polished surface.These curves of Figure 11 A curve map are represented the monitoring to infrared signal of on the diverse location of described band (all as described with the center front-back direction with respect to the operator).

Figure 11 B is the infrared signal time history plot according to the chip temperature of one embodiment of the invention.Here, the monitoring wafer temperature is so that monitor variation of temperature in CMP technological process process.For each embodiment of Figure 11 A and 11B, at the destination county of the process of association, line 218 and 220 intersects with the response curve of described curve map respectively.In other words, the temperature of band and the temperature of wafer begin to descend at destination county, and described terminal point is on described point of inflexion on a curve.In one embodiment, in described processing procedure, about 20 ℃ of wafer temperature variations, and about 10 ℃ of the temperature variation of band.

Figure 12 is the synoptic diagram of 30 seconds time series T1-T9 of CMP process, and graphic extension is according to the removal process of the copper film that utilizes the vortex sensor measuring in the chip carrier of one embodiment of the invention.Among the time series T1-T9 each all illustrates the ECS signal (volt) on the y axis and the time interval (second) on the x axle.Time series T1 illustrates the beginning of described process, and the removal process starting point of the copper of time series T2 explanation simultaneously film.In other words, remove 479  materials at time series T2.Time series T3-T8 illustrates the amount of removing of about 30 second cycle and material relevant in described time series process.The appearance of time series T9 explanation terminal state.

Figure 13 is the synoptic diagram of simplifying, and graphic extension is applied to or the dutycycle of the handover scheme of skew or coaxial two sensors according to one embodiment of the invention.Here, the dutycycle 250 that interrelates with first sensor staggers with respect to the dutycycle 252 that interrelates with second sensor.In other words, when dutycycle 252 is in " leading to " during state, dutycycle 250 is in, and " breaking, " state, vice versa.Like this, utilize this interleaved power scheme to eliminate these sensors by just in the cross-couplings of measured substrate.Should be pointed out that the interleaved power scheme can be called handover scheme.Will show in Figure 14 A to 16B that as following with all scheme of the power supply contrasts simultaneously of two sensors, the noise that Figure 13 handover scheme causes significantly reduces.Should be pointed out that first and second sensors can depart from each other, that is the axis of described second sensor of the axis runout of first sensor is as in Fig. 3,6A and 7B graphic extension.As another program, as in Fig. 5 graphic extension, first sensor and described second sensor can be coaxial under described configuration.

Figure 14 A and 14B are exemplary graph, and graphic extension is according to one embodiment of the invention, and non-switching power supply plan and the noise that switches between the power supply plan be poor.The voltage readings of the non-switching power supply plan of Figure 14 A graphic extension.The voltage readings of power supply plan is switched in Figure 14 B graphic extension.As can be seen, the noise level that interrelates with Figure 14 A is significantly greater than the noise level that interrelates with Figure 14 B.As mentioned above, the reduction of noise level is to eliminate the coupling by wafer of when they are switched on simultaneously first and second sensors to cause.Like this, by alternately being each power supply in the sensor, can obtain reading more accurately, less interference.Should be pointed out that it is these sensors power supplies that the embodiments described herein allows to come with single power supply, thereby eliminate the error that the different noisinesss by different electrical power cause.

Figure 15 A and 15B are exemplary plot, are illustrated in the noise that causes in the non-switching power supply plan.The zone 254 of exploded view 15A among Figure 15 B.Here, curve 256 representatives are from the signal of upper sensor, and curve 258 representatives simultaneously are from the lower sensor voltage signal.The curve 256 of expansion in the zone 254, as indicated in Figure 15 B, voltage readings vibration between 7.7 and 7.9 volts in general.Figure 16 A and 16B are exemplary plot, and representative is according to one embodiment of the invention, when switching the reading that is similar to Figure 15 A and 15B when power supply plan is applied to sensor.Curve 260 is represented the lower sensor voltage readings, and curve 262 is represented the upper sensor voltage readings simultaneously.As in Figure 16 A, seeing voltage readings and corresponding signal contrast relatively flat in Figure 15 A.Extended area 264 further graphic extension curve 260 relatively flats in Figure 16 B.Here, the voltage readings of lower sensor still is a quite stable, all in 7.75 and 7.8 volts zone, and forms sharp contrast between 7.7 and 7.9 volts of Figure 15 B.

Figure 17 is the synoptic diagram of simplifying, and graphic extension improves according to edge exclusion in the arranged coaxial of one embodiment of the invention.Wafer 280 comprises two concentric circless of with dashed lines 282 and 284 graphic extensions.In offset configuration, eddy current sensor can be limited to the thickness of measuring in the zone that is defined in the curve 284.But in arranged coaxial, described zone can be to be extended to the zone that is defined in the curve 282.Like this, here, measure the bigger scope of wafer 280.For example, be roughly under 18 millimeters the situation at the diameter of eddy current sensor probe, the area that can measure can enlarge about 9 millimeters again.Similarly, be that survey area can enlarge at least 6 millimeters approximately again under 12 millimeters the situation at the diameter of described probe.

Figure 18 is a process flow diagram, is used in the measured film thickness process of graphic extension according to one embodiment of the invention inspection spot definition and noise are reduced to minimum method operation.Described method, is arranged to first eddy current sensor towards the first surface related with conductive film in operation 270 from operating 270.Described method enters operation 272, in operation 272, second eddy current sensor is arranged to towards the second surface related with conductive film.Here, describe with reference to Fig. 3 and 5 as top, second eddy current sensor and first eddy current sensor can be aimed at the two opposite sides of Semiconductor substrate.Should be pointed out that first eddy current sensor and second eddy current sensor can be coaxial or depart from each other, as graphic extension here.When eddy current sensor when being coaxial,, reduced the inspection spot definition with the offset configuration contrast.For example, in offset configuration,, check that two probe diameter of spot definition and these eddy current sensors are the same big as reference Fig. 3 graphic extension.But, in arranged coaxial,, check that spot definition is reduced into the single diameter of eddy current sensor probe as reference Fig. 5 graphic extension.In addition, eddy current sensor can be positioned at such as the handling implement of chemical mechanical planarization process instrument or as with a part of aiming at the mapping function that station interrelates.

The method of Figure 18 enters operation 274 then, in operation 274, power supply alternately is connected to first eddy current sensor and second eddy current sensor.When in other words, the power supply of first eddy current sensor was " leading to ", the power supply of second eddy current sensor was " breaking ".Therefore, once only power supply in these eddy current sensors, thus eliminate the cross-couplings of these eddy current sensors by substrate.In one embodiment, can be being included in time delay in the power supply plan.Connection in other words, once and again " " and " disconnections " power supply of first eddy current sensor, at " connection " power supply of second eddy current sensor before through a time delay.In one embodiment of the invention, can be one millisecond described time delay, still, can apply any suitable time delay.Described method moves on to operation 276, in operation 276, calculates the measured film thickness value according to the combination from the signal of first eddy current sensor and second eddy current sensor.Because the interleaved power scheme makes noise reduce to minimum, so the film thickness of calculating will have higher accuracy and degree of accuracy.

In one embodiment of the invention, wherein first and second eddy current sensors are coaxial, and described relative eddy current sensor can be configured to occur with the form of the minimum inductance load of corresponding eddy current sensor.Those skilled in the art will recognize that this can realize by eddy current sensor is included in the open cycle system.In addition, the embodiments described herein allows with single power supply to first eddy current sensor and the power supply of second eddy current sensor, because power supply replaces.Thereby, got rid of the error that a plurality of power supplys are introduced when eddy current sensor is powered, for example, the noisiness of different power supplys.In addition, can get rid of the edge exclusion that in offset configuration, takes place by in arranged coaxial, using the interleaved power scheme.In other words, when sensor head during near the edge of wafer, the part of probe will be exposed at outside the wafer circumference, and a part is exposed on the circumference of wafer.This result who exposes to the open air is an edge exclusion, wherein measures the fringe region less than wafer.Described fringe region may be the same with the diameter of eddy current sensor big, thereby, in arranged coaxial and skew eddy current sensor contrast fringes repel and will reduce to minimum.In other words, the edge exclusion zone can be reduced into the radius of sensor.

Say blanketly, the invention provides by the sensor configuration of trooping and determine the method for semiconductor fabrication such as the CMP course end.Described troop sensor for by at first at the thickness of determining the film on the wafer under the state of not processing, determine that terminal point gets ready with the relevant removal or the speed of deposit.In one embodiment, just measured film thickness is between 0 micron and 2 microns.Described definite thickness is offered second sensor that interrelates with process operation, so that demarcate described sensor, the feasible variable that causes error in thickness measurements from processing conditions is excluded basically.Although should be pointed out that with regard to these embodiment of CMP process prescription, these sensor arraies of trooping are not limited to the CMP process.For example, described sensor can use on substrate and remove or layer of deposit or film, in any semiconductor processes such as etching and deposition process.In addition, defined the switching power supply plan that is used for minimum.Switch the coupling of these sensors of taking place when power supply plan is eliminated opposed sensor and powered simultaneously by substrate.When the form with arranged coaxial is provided with sensor, described switching power supply plan, and, reduced the degree of edge exclusion by sensor being comprised the minimization process that in open cycle system, makes the relative sensor that sensor occurs with the form of inductive load.

Here the present invention has been described with regard to several one exemplary embodiment.Consider instructions of the present invention and practice, other embodiments of the invention will become apparent the professional and technical personnel.Should regard the foregoing description and preferable feature as exemplary, the present invention is defined by the accompanying Claim book.

Claims

1. one kind is used in the measured film thickness process inspection spot definition and noise being reduced to minimum method, and described method comprises:

First eddy current sensor is arranged to first surface towards the substrate related with conductive film;

Second eddy current sensor is arranged to the second surface related with described conductive film towards described substrate, and described second surface is relative with described first surface;

Alternately to described first eddy current sensor and the power supply of described second eddy current sensor, make and when described first eddy current sensor is powered, do not power, and when described first eddy current sensor is powered, do not power to described second eddy current sensor to described second eddy current sensor; And

According to the described measured film thickness value of combination calculation from the signal of described first eddy current sensor and described second eddy current sensor.

2. the method for claim 1, wherein said method is arranged to second eddy current sensor to comprise towards the operation of the second surface related with described conductive film the axis of described second eddy current sensor of the axis runout that makes described first eddy current sensor.

3. method as claimed in claim 2, wherein said method alternately comprise to the operation of described first eddy current sensor and the power supply of described second eddy current sensor:

A) to described first eddy current sensor power supply;

B) end is to the power supply of described first eddy current sensor;

C) wait for one period time delay;

D) to described second eddy current sensor power supply;

E) end is to the power supply of described second eddy current sensor; And

F) wait for one period described time delay.

4. method as claimed in claim 3 wherein also comprises: for each measuring position repeating step a)-f).

5. the method for claim 1, wherein said method is arranged to second eddy current sensor to comprise towards the operation of the second surface related with described conductive film: aim at described first eddy current sensor, make it coaxial with described second eddy current sensor.

6. method as claimed in claim 5, wherein said method alternately comprise to the operation of described first eddy current sensor and described second eddy current sensor power supply: described first eddy current sensor and described second eddy current sensor are configured to when corresponding eddy current sensor is inoperative form appearance with minimized inductive load.

7. method as claimed in claim 6, wherein said method are configured to when corresponding eddy current sensor is inoperative the operation that form with minimized inductive load occurs to described first eddy current sensor and described second eddy current sensor and comprise: described first eddy current sensor and the described second eddy current sensor both are included in the open loop.

8. the method for claim 1 wherein also comprises:

Power to described first eddy current sensor and described second eddy current sensor from single power supply; And

Repeat the described blocked operation of described power supply, so as for each position the time that replaces to described first and described second eddy current sensor power supply.

9. a sensor array that is used to shine upon wafer thickness comprises:

A plurality of top sensor;

The a plurality of bottom sensor relative with described top sensor, in in wherein said a plurality of bottom sensor each and the described a plurality of top sensor corresponding one coaxial, described a plurality of bottom sensor also be configured to when in described a plurality of top sensor corresponding one inoperative when working;

The power supply of keeping in touch with described a plurality of top sensor and described a plurality of bottom sensor; And

Controller is configured to alternately power to described a plurality of bottom sensor and described a plurality of top sensor from described power supply.

10. sensor array as claimed in claim 9, wherein said a plurality of top sensor and described a plurality of bottom sensor all are eddy current sensors.

11. sensor array as claimed in claim 9, wherein said controller also are configured to when described sensor is comprised one period time delay when inoperative state switches to the state that works.

12. sensor array as claimed in claim 9, wherein said sensor array are included in the aligning station of semiconductor processes instrument.

13. sensor array as claimed in claim 11, be 1 millisecond wherein said time delay.

14. a system that is used to handle wafer comprises:

Chemical-mechanical planarization (CMP) instrument, described CMP instrument comprise,

Be defined chip carrier in the enclosure, the bottom surface of described chip carrier has the window that is limited on this bottom surface;

Be attached to the carrier thin film on the described bottom surface of described chip carrier, described carrier thin film is configured to support wafer in the CMP operating process; And

Embed the sensor in the described wafer, described sensor is arranged on above the end face of described window, and described sensor configuration becomes at the described wafer eddy current of inducting, so that determine the degree of approach and the thickness of described wafer;

Sensor array beyond described CMP instrument, described sensor array and the sensor communication that embeds described chip carrier, described sensor array comprises first sensor and corresponding second sensor, described first sensor and corresponding second sensor configuration strike a bargain and are in state and the inoperative state that works for ground, described first sensor also is configured to be in the state that works when described second sensor is in inoperative state, described sensor array be configured to detect wafer with described first sensor and corresponding second sensor wafer thickness signal to the range-independence of described wafer.

15. system as claimed in claim 14, the described first sensor of wherein said sensor array and corresponding second sensor have common axis.

16. system as claimed in claim 14, the axis of corresponding second sensor of the axis runout of the described first sensor of wherein said sensor array.

17. system as claimed in claim 16 wherein asks from the signal of described first sensor and mean value from the signal of corresponding second sensor, so that determine original depth.

18. system as claimed in claim 14 wherein also comprises:

The power supply of keeping in touch with described first sensor and corresponding second sensor; And

Controller, it is configured to alternately to described first sensor and the power supply of corresponding second sensor.

19. system as claimed in claim 15, wherein said controller is communicated by letter with the sensor and the described sensor array of described embedding, described controller is configured to determine according to the signal that sensor or described sensor array by described embedding provide the thickness of described wafer, and described controller can be to the wafer before the CMP controller provides CMP and the thickness distribution curve of the wafer after the CMP.